While tech giants continue to pour hundreds of billions of dollars into AI hardware infrastructure, recent data from Goldman Sachs, Apollo, and MIT collectively reveals a troubling trend: enterprise AI adoption is decelerating, with large corporations experiencing declining adoption rates.
According to a September 8 report by Goldman Sachs Chief Economist Jan Hatzius, despite continued acceleration in AI hardware investment, the pace of AI adoption among US enterprises has slowed in the third quarter.
Data shows that in Q3 2025, the percentage of US companies using AI only marginally increased from 9.2% in Q2 to 9.7%. While still growing, the growth rate has decelerated. This stands in stark contrast to the heated investment in AI infrastructure, where tech giants have invested hundreds of billions of dollars in AI data center construction—massive capital expenditures that urgently need to demonstrate return on investment.
The report also indicates that financial and real estate sectors showed the largest increases in AI adoption, while educational services experienced a decline. Regarding the labor market, Goldman Sachs considers the impact remains "moderate," though job displacement has emerged in technology, design, and customer service sectors. Since the last update, AI-related layoffs have affected 10,375 workers.
**More Concerning Signal: Large Enterprise AI Adoption Rates Are Declining**
More worrying than Goldman Sachs' revelation of "slowing growth" is that large enterprises' AI adoption rates may have already peaked and begun declining.
According to analysis by Apollo Global Management Chief Economist Torsten Sløk, AI adoption among large enterprises is decelerating. Based on US Census Bureau survey data covering 1.2 million businesses, the analysis shows that companies with 250 or more employees are experiencing a downward trend in AI adoption rates.
This trend may signal that after an initial period of enthusiastic experimentation, large enterprises are entering a "trough of disillusionment," beginning to reassess the actual value and integration challenges of AI tools.
**95% of AI Investments Yielding Zero Returns?**
The fundamental reason behind enterprises' AI adoption obstacles may be found in an earlier MIT report that triggered market selloffs. A report titled "The Generative AI Gap: 2025 State of Business AI" found that up to 95% of enterprises are receiving zero returns from their generative AI investments.
Lead author Aditya Challapally noted that the core issue is not the AI models themselves, but rather "learning gaps" and integration strategy flaws within enterprises. For example, many general-purpose tools designed for individual users (such as ChatGPT) perform poorly in enterprise environments that require adaptation to specific workflows. Additionally, the report found that "buying" mature AI tools has a much higher success rate (approximately 67%) compared to enterprises "building" their own systems (about one-third), posing a direct challenge to companies that have invested heavily in developing proprietary AI systems.
**Market Panic May Spread**
These negative data points have already substantially impacted market sentiment. After the MIT report was released in August, it triggered significant tech stock selloffs, with the Nasdaq index recording its largest single-day monthly decline (-1.4%) and NVIDIA shares falling 3.5%. A trader commented at the time, "This story is making people panic."
For investors, from Goldman Sachs' "slowing growth" to Apollo's "declining adoption rates," and MIT's revelation of the "zero returns" predicament, this series of signals clearly indicates that AI's commercialization path is far more complex and lengthy than anticipated. With the Nasdaq 100's forward P/E ratio still nearly one-third higher than its long-term average, investors need to shift their focus from frenzied pursuit of technological breakthroughs to prudent assessment of enterprises' actual implementation capabilities and profitability.